The Safety of Concurrent Administration of Opioids via Epidural and Intravenous Routes for Post-operative Pain in Pediatric Oncology Patients

Abstract

Supplementation of epidural opioid analgesia with intravenous opioids is usually avoided because of concern about respiratory depression. However, the choice of adjunct analgesic agents for pediatric oncology patients is limited. Antipyretic drugs may mask fever in neutropenic patients, and non-steroidal anti-inflammatory agents may exert anti-platelet effects and interact with chemotherapeutic agents. We examined the safety of concurrent use of epidural and intravenous opioids in a consecutive series of 117 epidural infusions in pediatric patients and compared our findings to those reported by other investigators. We observed a 0.85% rate of clinically significant respiratory complications. The single adverse event was associated with an error in dosage. In our experience, the supplementation of epidural opioid analgesia with intravenous opioids has been a safe method of postoperative pain control for pediatric patients with cancer.

Introduction

The combination of opioids and local anesthetics has become the standard of care for epidural analgesia, because its analgesic effects are greater, and its adverse effects are less, than those of local anesthetics alone (1). However, administration of epidural opioids incurs the risk of respiratory depression and sedation. Epidural analgesia is typically supplemented with non-opioid analgesic agents such as acetaminophen and non-steroidal anti-inflammatory drugs; intravenous opioids are usually avoided because of early reports of an increased incidence of respiratory depression (2).

The choice of adjunct analgesic agents for pediatric cancer patients is limited. Antipyretic agents may mask fever in neutropenic patients, and non-steroidal anti-inflammatory drugs can exert an anti-platelet effect and interact with chemotherapeutic agents (3). Consequently, epidural analgesia is frequently supplemented with intravenous opioids at our institution. Here we report the safety of this practice in our institution’s recent experience and discuss our findings in the context of the available literature.

Methods

Patients

This study was approved by the St. Jude Children’s Research Hospital Institutional Review Board, which waived informed consent. Data had been prospectively collected on all patients consecutively treated for postoperative pain by continuous epidural infusion between September 2004 and July 2006 as part of a Quality Improvement (QI) project led by the Pain Management Service. We identified all cases in which opioids were administered concurrently by epidural and intravenous routes. Patients were excluded from analysis if they were older than 18 years or if they received opioids by only one route.

Pain Management

Postoperative analgesia was provided by continuous epidural infusion of bupivacaine 0.1%–0.125% and fentanyl 2–5 mcg/ml, at an hourly rate calculated to deliver 0.5–1 mcg/kg of fentanyl per hour. Orders for supplemental analgesia consisted of nurse administered intravenous doses of opioid as needed (IV PRN). Dose ranges for supplemental analgesia were 0.05–0.1 mg/kg of morphine, 0.5–1 mcg/kg of fentanyl and 0.01–0.02 mg/kg hydromorphone every 1–2 hours as needed. If patients had a history of opioid tolerance and difficult-to-control pain, or if IV PRN analgesia failed to control pain, opioid IV patient-controlled analgesia (IV PCA) was provided. The first-line intravenous opioid was morphine; fentanyl and hydromorphone were used as alternatives. The recommended starting PCA bolus doses were: morphine 0.02 mg/kg, hydromorphone 0.004 mg/kg, and fentanyl 0.5 mcg/kg, with a 15-minute lockout interval. If a background infusion was added, the starting dose per hour was equal to the bolus dose.

Patients were admitted to the general oncology floor or the intensive care unit depending on the postoperative status, as decided by the anesthesiologist and the surgeon. The presence of an epidural catheter did not influence admission to a higher acuity floor.

Vital signs (blood pressure, heart rate, respiratory rate, and temperature) were monitored every four hours (every two hours in the intensive care unit). Pain intensity, level of consciousness, and motor function were monitored every four hours while patients were awake, and pulse oximetry was monitored continuously. The patient’s level of consciousness was categorized as alert, drowsy, confused, asleep, or unarousable. Pulse oximetry values were displayed at the nursing station via a central monitoring system. It is our hospital policy to notify the physician of oxygen saturations <95%, which may prompt an order for supplemental oxygen. For oxygen saturation <90%, there are standing orders to administer supplemental oxygen by simple face mask at 5 liters/minute.

Data Collection

The medical records of each patient receiving continuous epidural analgesia were examined daily, and relevant information was stored in the QI database. These data included the name and concentration of opioid and local anesthetics, the rate of epidural infusion, and any adjustments to those parameters. Data collected on supplemental intravenous opioid analgesia (IV PRN or IV PCA) included the drug, dose, and time interval. If the method of opioid administration (IV PRN vs. IV PCA) changed during the 3 to 4 days of concomitant continuous epidural infusion, we counted each patient in the category in which he was placed initially (as IV PRN or PCA). Each patient’s pain score (median and range) was summarized daily in the QI database. Age-appropriate pain scales were used: the numeric pain scale (NPS) for children older than 13 years, the FACES pain scale for those 5–13 years of age, and the FLACC behavioral scale for those younger than 5 years. Any change in respiratory status, including decreased respiratory rate, respiratory amplitude, or pulse oximetry values, was recorded in the database, as were any neurological changes, such as sedation, confusion, hallucination, or seizure. Corrective interventions were also recorded. We examined any adjustments in the epidural infusion rates or composition of the epidural infusate prompted by changes in respiratory or neurological status.

Definition of Adverse Events

If neurological or respiratory changes were noted, we examined each incident to determine its clinical significance and identify possible contributory factors, such as significant comorbidity or additional sedative drugs (e.g., benzodiazepines and phenothiazines). Clinically significant episodes of respiratory depression were defined as any occurrence of apnea or requirement for aggressive intervention (naloxone administration or intubation) in addition to discontinuation of opioids. Minor respiratory events were defined as a respiratory rate lower than ten breaths per minute and/or oxygen saturation less than 90% on room air, requiring minimal interventions such as stimulation, supplemental oxygen, and adjustment of the epidural infusion rate or substitution of PRN opioids for PCA opioids.

Results

Patients

One hundred and forty oncology patients received postoperative epidural analgesia after thoracotomy, exploratory laparotomy, and lower-extremity surgery. Twenty patients were excluded because of age >18 years and three patients were excluded because they received opioids by only one route. Therefore, we examined 117 consecutive cases of postoperative pain management given by continuous epidural opioid infusion and concomitant intravenous opioid infusion. Thirty patients (25.6%) were 1–2 years, 35 patients (30%) were 3–10 years, and 52 patients (44.4%) were 11–18 years of age. The mean age was 8 years (range, 8 months to 18 years). The intravenous opioids were administered initially as IV PRN doses in 103 cases and by IV PCA in 14 cases. The patients were categorized on the basis of initial therapy, but ten “crossed-over” (nine from IV PRN to IV PCA, and one from IV PCA to IV PRN) during the course of the continuous epidural infusion, which is typically three days. In the PCA group, five patients had basal opioid infusions augmented by patient-controlled boost doses, and nine had patient-controlled doses only. The anatomic level of catheter placement and the types of surgical intervention are presented in Table 1. The dermatomal level of the block is not routinely assessed, and was not reported.

Table 1.

Type of Surgery and Location of Epidural Catheter

Level of catheter placement (n)	Type of surgery
	Thoracotomy	Exploratory laparotomy	Lower extremity surgery
Thoracic (32)	18	14	0
Lumbar (84)	6	32	46
Caudal (1)	0	1	0
Total (117)	24	47	46

Efficacy of Analgesia

Pain was scored as mild (pain score 0–3), moderate (4–6), and severe (>7) on a 1–10 scale. The distribution of pain scores on postoperative days 1 to 3 is shown in Table 2.

Table 2.

Distribution of Pain Scores on Postoperative Days 0 to 4

	Maximum Pain Scores			Median Pain Scores
	0 to 3	4 to 6	7 to 10	0 to 3	4 to 6	7 to 10
Day 0 (n = 101)	42 (41.6%)	34 (33.7%)	25 (24.7%)	69 (68.3%)	22 (21.8%)	10 (9.9%)
Day 1 (n = 108)	46 (42.6%)	39 ( 36.1%)	23 (21.3%)	83 (76.9%)	15 (13.9%)	10 (9.2%)
Day 2 (n = 104)	59 (56.7%)	26 (25%)	19 (18.3%)	85 (81.7%)	14 (13.5%)	5 (4.8%)
Day 3 (n = 67)	43 (64.1%)	17 (25.4%)	7 (10.5%)	55 (82.1%)	10 (14.9%)	2 (3%)

Adverse Events

One clinically significant respiratory adverse event was recorded (0.85%). A 10-month-old infant experienced recurrent apnea and bradycardia after inadvertent administration of 5 mg of morphine rather than the 0.5 mg prescribed, while receiving supplemental oxygen. This complication resolved with stimulation, supplemental oxygen, and discontinuation of the epidural infusion. Minor respiratory events were observed in two patients, 14 and 17 years of age, who experienced respiratory rates of 7 and 9 breaths per minute, respectively. Neither patient required supplemental oxygen and were managed by substituting IV PRN morphine for the IV PCA in the former case and by decreasing the rate of epidural infusion in the latter. Naloxone was not required for any episode of opioid-mediated respiratory depression and neither route of opioid administration had to be discontinued.

Safety Analysis

No neurological complications were observed. Of the three respiratory complications observed, two (including the drug error described above) occurred in the IV PRN opioid group and one occurred in the IV PCA group in a patient receiving a basal opioid infusion. Only one respiratory adverse event was clinically significant, and was noted in the IV PRN opioid group.

We also examined the records of the three patients who received opioids via only one route. Two of these patients experienced adverse respiratory events. One patient receiving epidural fentanyl and bupivacaine after a thoracotomy had arterial oxygen desaturation to 75% while receiving supplemental oxygen and had a pleural effusion. The other patient had a respiratory rate of five breaths per minute while receiving IV PCA opioid and epidural analgesia with local anesthetic only; his lowest oxygen saturation was 94% and he did not receive supplemental oxygen.

Epidural infusion rates were decreased in six cases (6 of 117). Two cases are presented above, and a further four cases did not meet the study criteria for either clinically significant or minor respiratory events. No adjustments to the concentration of opioid in the epidural infusate were prompted by respiratory or neurological changes; rather they were prompted by pruritis, hypotension or bradycardia.

Discussion

Only one of the 117 consecutive pediatric patients who received concomitant epidural and intravenous opioids (0.85%) had a clinically significant respiratory complication, which was the result of a drug error. Two other respiratory events were minor and did not result in discontinuation of the epidural or intravenous opioids.

Continuous epidural analgesia (bupivacaine 0.1%–0.125% and fentanyl 2–5 mcg/ml for 72 hours) is the standard postoperative practice at our institution. Although epidural opioid analgesia is not ordinarily supplemented with IV opioids, it is used for our pediatric oncology patients for several reasons. First, the choice of agents is limited. Our clinicians are concerned that acetaminophen may mask fever in neutropenic patients. Nonsteroidal anti-inflammatory drugs are contraindicated for patients with thrombocytopenia and they interact with chemotherapeutic agents. Second, most of our patients require chronic opioid pain control and have some degree of opioid tolerance at the time of surgery.

The high rate of supplemental parenteral opioid analgesia in this study is not necessarily a reflection of the failure of the epidural analgesia, rather an institutional approach to provide liberal “back up “ analgesia plans for all patients receiving regional or neuraxial analgesia. The first line intravenous opioid supplementation in our practice is PRN administration of IV opioid doses. The rationale for choosing IV PRN doses as the preferred modality is that our low nurse to patient ratio of 1:2 allows frequent clinical observation and prompt provision of PRN opioids as necessary. The PRN doses of opioids are ordered with a range of doses, the higher dose being one that would be administered in the absence of epidural opioids as a “full dose” and the lower dose being a “half dose.” The nurses frequently choose the lower dose first, and then escalate the dose as necessary. IV PCA is used for selected patients with a history of opioid tolerance and pain that is difficult to manage. IV PCA opioids can also be a contingency strategy for patients whose pain is not adequately controlled by intermittent IV doses. However, because one of our 14 patients receiving IV PCA with a basal infusion experienced a minor respiratory complication, basal infusion is no longer used in conjunction with IV PCA and epidural opioids, as it may carry a risk of complications. Patients were not randomized to either IV PRN or IV PCA groups, and crossover occurred between the two groups. For both of these reasons no conclusion can be drawn regarding the safety of one rescue technique over the other.

Data about epidural opioid-induced respiratory depression in children are limited. Overall, the reported incidence of respiratory depression associated with epidural opioids is 0% to 1.9% in adults (Table 3) (1,2,4–19) and 0–25% in children (Table 4) (20–30). The rates most commonly reported in children are 0.5%–1% (24).

Table 3.

Incidence of Respiratory Depression with Epidural Analgesia – Adult Data

Study, Year, Ref #(n)	Study Design	Epidural Opioid	Incidence of Respiratory Depression	Criteria
Wheatley, 2001(1) (1014 – > 1.3 million)	Review of RCTs	various	0.24–1.6%	Various
Wheeler, 2002(19) (1596)	Review of RCTs	various	E: 1.9% IV PCA: 1.8% IV/IM: 2.4%	Various
Fischer, 1988 (37) (107)	Prospective	morphine fentanyl	0%	“Clinically evident”
Ready, 1991(9) (1106)	Prospective	morphine	0.2%	Naloxone use
Stuart-Taylor, 1992(13) (800)	Prospective	diamorphine	0.9%	RR < 10/min
de Leon-Casasola, 1994a (4) (4227)	Prospective	morphine	0.07%	RR < 10/min prompted use of naloxone per protocol.
Scott, 1995(11)(1014)	Prospective	fentanyl	RR<8/min: 1.2%, Naloxone: 0.4%	RR < 8 Naloxone use
Rygnestad, 1997 (10) (2000)	Prospective	morphine	Overall: 1.6% Severe: 0.15%	Considered severe if infusion was stopped.
Liu, 1998b (8)(1030)	Prospective	fentanyl	0.3%	RR < 8/min
Flisberg, 2003(5) (2696)	Prospective	morphine	IV PCA: 1.2% PCEA: 0.4%	RR < 8/min
Shapiro, 2005(12) (1524)	Prospective	morphine	Overall: 1.2% IV PCA: 1.9% E: 0.6% IT: 0.7%	RR < 10/min
Burstal, 1998(38) (1062)	Prospective survey	various	0.32%	RR < 8 and sedation
Fuller, 1990(6) (4880)	Retrospective	morphine	0.25%	RR < 10/min
Tsui, 1997(14) (1466)	Retrospective	morphine fentanyl	IV PCA: 1.97% E: 0.6%	RR < 10/min SpO2 < 90% PCO2 > 7kPa
Wigfull, 2001b (16) (1057)	Retrospective	fentanyl	0.19%	RR < 8/min

RCT = randomized controlled trial; E = epidural; IV = intravenous; PCA = patient controlled analgesia; IM = intramuscular; RR = respiratory rate; PCEA = patient controlled epidural analgesia; IT = intrathecal; SpO₂ = arterial oxygen saturation; pCO₂ = arterial carbon dioxide partial pressure.

^aUsed intravenous morphine as supplementary analgesia.

^bUsed patient-controlled epidural analgesia.

AU: PLS CHECK SUPERSCRIPTS AS THEY HAD TO BE CHANGED TO CONFORM TO JOURNAL STYLE. ALSO, IS THE ADDITION OF REF # TO COLUMN 1 OKAY WITH YOU?

Table 4.

Incidence of Respiratory Depression with Epidural Analgesia – Pediatric Data

Study, Year, Ref #(n)	Study Design	Opioid	Incidence of Respiratory Depression	Criteria
Krane, 1989(26) (32)	RCT	morphine	3.1%	Not specified
Kart, 1997a(31)	RCT	fentanyl morphine	0%	Not specified
Goodarzi, 1999 (23) (90)	RCT	morphine fentanyl hydro-morphone	morphine: 25%(none severe) fentanyl: 0% hydromorphone: 0%	RR < 10/min, SpO2 < 90%
Attia, 1986(20) (20)	Prospective	morphine	0%	Apnea
Krane, 1987(25) (46)	Prospective	morphine	0%	RR<10/min
Lovstad, 1997(28) (100)	Prospective	fentanyl	0%	Treatment with rescue medication
Giaufre, 1996b (22) (15,013)	Prospective survey	various	0.006%	Apnea
Shayevitz, 1996(30) (54)	Retrospective case control	morphine	E: 0% IV: 0%	Delayed extubation, reintubation, naloxone use
Valley, 1991(29) (138)	Retrospective	morphine	8%	Apnea, desaturation, bradycardia, decreased RR, treated with stimulation, intubation, or nalaxone
Flandin- Blety, 1995(21) (7200)	Retrospective questionnaire	various	0%	Various, not specified
Williams, 2003(24)	Retrospective questionnaire	various	Overall: 0.1–5%	Various, not specified

RCT = randomized controlled trial; IV = intravenous; RR = respiratory rate; SpO₂ = arterial oxygen saturation; E = epidural.

^aUsed intravenous morphine as supplementary analgesic.

^bUsed mostly caudal blocks.

AU: PLS CHECK SUPERSCRIPTS AS THEY HAD TO BE CHANGED TO CONFORM TO JOURNAL STYLE. ALSO, IS THE ADDITION OF REF # TO COLUMN 1 OKAY WITH YOU?

The lack of consistent reporting criteria limits the meaningful comparison of studies about neuraxial opioids. Some investigators report respiratory depression on the basis of requirement for naloxone reversal (2,9,11,31), indicating a relatively severe complication. When this definitive criterion is used, the reported rate of respiratory depression with epidural opioids is 0.2–0.4% (2,9,11). However, when reduced respiratory rate and oxygen saturation are included as criteria, rates of 0.07 %–1.2% are reported in adults (Table 3) (4–6,8,11–14,16). Similarly, pediatric studies of epidural analgesia have used two broad definitions of respiratory depression: 1) clinically significant episodes such as re-intubation, use of naloxone or apnea (20,22,29,30); and 2) definitions that include minor events such as respiratory rate <10 per minute or oxygen saturation < 90% (23,25). One potential criticism of our study is that it did not use age specific respiratory rates, which may be relevant in small children and babies. Our definitions of adverse respiratory events were designed to be as inclusive as possible, whilst maintaining comparability to other studies.

The respiratory depression rates associated with epidural and intravenous opioids have been compared. While morphine administered epidurally appeared to be associated with a lower rate of respiratory depression (0.4%–0.6%) than IV PCA morphine (1.2%–1.9%) (5,12), a systematic review of randomized controlled studies found comparable cumulative respiratory depression rates of 1.9% for epidural opioids and 1.8% for IV PCA (19).

In the case of PCA opioid administration in adults respiratory depression rates of 1.2% to 11.5% have been found with meta-analysis and 1.8 % in a systematic review of randomized controlled trials. In children the incidence of respiratory depression associated with the use of PCA opioids is 0%–7% (32–35) and at our institution 0.56% (32).

The respiratory depression rate (0.85%) found in our investigation of dual-route opioids is comparable to that reported for single-route epidural opioid administration or PCA in adults and children. It is also comparable to our institutional complication rate of 0.56% with PCA opioids alone (32). This result indicates an acceptable level of safety. However, we acknowledge that a second route of opioid administration introduces an additional venue for error. The only clinically significant respiratory event in our study was caused by a drug overdose due to human error.

One investigator has reported routine concurrent administration of IV and epidural opioids in adult oncology patients (4). Intravenous morphine given concurrently with epidurally infused bupivacaine and morphine for postoperative pain was associated with a 0.07% rate of respiratory depression requiring reversal with naloxone. This low incidence of respiratory complications, as in our patient population, may reflect the increased likelihood of opioid tolerance caused by opioid administration for chronic cancer pain (36).

There are few reported studies of concurrent opioid administration via two routes in non-oncology patient populations. A pediatric study comparing continuous epidural infusion of fentanyl and bupivacaine versus intermittent epidural morphine for postoperative pain control found no episodes of respiratory depression, despite the concomitant use of IV opioids in 7% and 40% of cases, respectively (27).

In a landmark study, Gustafsson et al. identified concomitant use of intravenous opioids or sedatives, residual anesthesia, supine position, and thoracic level of the epidural infusion as risk factors for respiratory depression with epidural opioids (2). Other investigators established a direct correlation between intraoperative IV fentanyl and postoperative respiratory depression when epidural opioids were given postoperatively (12).

Our observation of respiratory complications in two of the three patients who received opioids via only one route (epidurally or intravenously) suggests that patients’ overall medical condition, comorbid conditions, and other individual factors may contribute more significantly to the risk of respiratory complications than the concurrent administration of opioids by epidural and intravenous routes.

It has been often recommended that sedative-hypnotic drugs not be used together with neuraxial opioids as it may increase the risk of respiratory depression. Two minor respiratory events in our series were in cases in which systemic opioids were combined with additional CNS depressant medications (an antihistaminic and a benzodiazepine, respectively). However, to place this in context, the use of benzodiazepines and sedative antihistamines in our institution is almost universal.

We believe that the opioid tolerance of our patients increases the margin of safety of dual-route opioids and that our results cannot necessarily be extrapolated to a more general, opioid-naïve patient population. Moreover, our patients and their families tend to be medically sophisticated because of their prolonged hospital stays and cancer treatment and may be more alert to subtle changes. Other institutional factors may increase the safety of dual-route opioid administration. Our nurse-to-patient ratio for patients with epidural catheters rarely exceeds one to two, and all nurses complete an epidural competency program that emphasizes the need for careful monitoring. Further, patients are monitored continuously by pulse oximetry which is connected to a central monitoring system with alarms at the nursing station. These advantages may not be available in all institutions.

Our data suggest that in an appropriate environment it is safe to use intravenous opioids, given either PRN or via PCA, to supplement postoperative opioid-containing epidural anesthesia in pediatric oncology patients.